PREVIOUS MODELS OF THE IMPRISONMENT PROCESS HAVE STATED THAT THE NUMBER OF CRIMINALS AT LARGE IN SOCIETY IS REGULATED BY THE IMPRISONMENT OF CRIMINAL LAW VIOLATORS. WHEN THE NUMBER OF VIOLATORS IN PRISON DECLINES, THE NUMBER AT LARGE INCREASES. SOCIETY RESPONDS TO THIS INCREASE WITH MORE INCARCERATIONS, REDUCING THE NUMBER OF CRIMINALS PRESENT OUTSIDE THE PRISON SYSTEM. ONE SUCH MODEL SUGGESTS THAT THE MEAN SIZE OF THE CRIMINAL POPULATION IS NOT VERY RESPONSIVE TO SHORT-TERM CHANGES IN POLICY CONCERNING PRISON ADMISSIONS AND RELEASES, BUT IT DOES ATTRIBUTE CHANGES IN THE SIZE OF THE CRIMINAL POPULATION AT LARGE TO VARIATION IN THE SIZE OF THE PRISON POPULATION. THE MODEL PROPOSED CONSIDERING RELEASES FROM PRISON, THEN ADMISSIONS. THE FIRST ASSUMPTION ABOUT RELEASES THAT WAS TESTED WAS THAT THE NUMBER OF PRISONERS RELEASED IS PROPORTIONAL TO THE PRISON POPULATION IN ANY GIVEN YEAR. A MATHEMATICAL ANALYSIS OF THE ASSUMPTION SHOWS INSTEAD THAT THE LENGTHS OF THE SENTENCES SERVED IN PRISON ARE GEOMETRICALLY DISTRIBUTED. THE MODEL THEN TESTS THE ASSUMPTION THAT ADMISSIONS FLUCTUATE RANDOMLY ABOUT A CONSTANT LEVEL. THE OSCILLATORY BEHAVIOR OF PRISON ADMISSIONS IS RELATED TO SOME VARIABLE WHICH INFLUENCES THEM AND WHICH ALSO OSCILLATES WITH AN APPROPRIATE FREQUENCY. IT WAS FOUND THAT THOSE VARIABLES WHICH CHARACTERIZE THE STATE OF THE BUSINESS CYCLE SHOW THE REQUIRED BEHAVIOR. IT IS SUGGESTED THAT WHEN UNEMPLOYMENT INCREASES, THE CRIME RATE INCREASES. JUDGES RESPOND TO THE INCREASE BY IMPOSING PRISON SENTENCES MORE FREQUENTLY, SO PRISON ADMISSIONS INCREASE. AN ANALYSIS OF TRENDS IN THE TWO VARIABLES, UNEMPLOYMENT RATES AND PRISON ADMISSIONS, SHOWS A CLOSE CORRELATION BETWEEN THE TWO. THE MATHEMATICAL EQUATIONS USED TO TEST THE MODEL'S ASSUMPTIONS ARE GIVEN. (VDA)